Automated Synthesis and Optimization of Robot Configurations

Abstract

We present an extensible system for synthesizing and optimizing robot configurations. The system uses a flexible representation for robot configurations based on parameterized modules; this allows us to synthesize mobile and fixed-base robots, including robots with multiple or branching manipulators and free-flying robots. Synthesis of modular robots is also possible with our representation. We use an optimization algorithm based on genetic programming. A distributed architecture is used to spread heavy computational loads across multiple workstations. We take a task-oriented approach to synthesis in which robots are evaluated on a designer-specified task in simulation; flexible planning and control algorithms are thus required so that a wide variety of robots can be evaluated. Our system’s extensibility stems from an object-oriented software architecture that allows new modules, metrics, controllers, and tasks to be easily added. We present two example synthesis tasks: synthesis of a robotic material handler, and synthesis of an antenna pointing system for a mobile robot. We analyze several key issues raised by the experiments and show several important ways in which the system can be extended and improved.

Emergent Hand Morphology and Control from Optimizing Robust Grasps of Diverse Objects

“…Optimization-based methods have been used for automating the co-design of morphology and control. These include gradient-free methods, such as evolutionary algorithms [17]- [19], and gradient-based methods, such as constrained optimization [20], [21], differentiable simulation [22], [23], and policy optimization [24]. However, these methods require either great amount of samples to search in the high-dimensional design space or manual specification of constraints.…”

Optimizing a Quadruped Robot: A Comparison of Two Methods

“…Robots can evolve controllers for their locomotion, environment specific behaviors, their morphology, or multiple things at once. Several papers used GA to optimize the morphology of quadrupedal machines, mostly virtual creatures [Larpin 2011][Heinen 2009][Bongard 2011][ Nygaard 2016][Leger 1999]. Even though the use of GA and more broadly methods of evolutionary computing have been around for decades we can still consider their use in everyday engineering as an experimental approach.…”

Designing actuation systems for animatronic figures via globally optimal discrete search

“…Prior work has tackled design problems for robotic systems with multiple actuated degrees of freedom, typically in the context of locomotion or flight [Du et al 2016;Geilinger et al 2018;Jelisavcic et al 2017;Leger et al 1999;Megaro et al 2015]. Some of these works formulate the design task as tree or graph search problems [Ha et al 2018a;Zhao et al 2020], as we do.…”